Yoshi P talks about Warriors- from Producer Letter Live @ TGS

[Zanther]

Not to mention, 8man dungeons other than primal fights tend to have alot of adds appearing when engaging boss fight.. This is where WAR will shine alot!

I can count on 1 hand how many packs in coil have more than have more than 2 or 3 adds. Turn 4 is the only boss with adds that need to be tanked, and half of them 2 shot warriors.

[Churchill]

And the other half are more easily tanked by a Paladin who can regen their MP in combat so they aren't running out of TP by spamming Overpower.

[Onisake]

There would have to be a MAJOR oversight on our part (as players) for WAR to suddenly become on par with PLD. Given the amount of research we have already done, I find this unlikely.

1) if it's taken the entire playerbase this long to figure out, it's too well hidden. Therefore it is a design problem.

2) Perception is king. even if, on paper or in the data, they are roughly equal it won't matter. people feel WAR is weaker. therefore it is weaker.

3) The main comparison is performance in coil. Coil wasn't really designed to be done with the gear currently available pre-coil. so it's technically off-label use. If this is the case, than the issue isn't with WAR being weak, it is with PLD being too strong. it's able to do things it shouldn't be able to do. IE: consistenly tank mobs that are 10+ levels above your current iLvl.

-----------------------

There are a couple of things I don't like about WAR currently. I mentioned briefly a few things that could help. but the problem is far more deeply rooted than that.

both tanks have a 'mitigation' but it depends on your viewpoint. as far as an overall 'system' we can look at it like this:

HP is a resource. Damage removes the resource. Healing restores that resource.

A mob does damage. the tank takes damage. the healer restores damage.

Boiled down, Mitigation is the capability of the tank to restore 'free' HP to itself. This can be in the form of reducing the damage a mob does (PLD) or restored health (WAR). 'true' mitigation, however is really just reducing damage taken (PLD).

If we wanted to make paladin more on par with Warrior, Shield oath would need to be recast every 30 seconds. Spirits within, Cover, and Hallowed ground would need to remove shield oath, cost MP, and you could only apply shield oath after 2 Halone combos. to top it all off, Shield oath would need to constantly drain MP, similar to a bard song. (forcing you to work riot blade into your rotation) Sound stupid? you're right. it is.

but Why so severe?

looking back at the system for damage taken/restored. Paladin does its mitigation up front. War does its mitigation on the backend. Shield oath provides a persistent 20% mitigation as PLDs main form of mitigation. War, on the other hand does it on the back end. The difference with WAR is that it's a 15% to healing recieved, rather than a straight 20% from damage taken. However given that you must spend wrath stacks to use your secondary form of mitigation (IB, etc.), you lose the 15% temporarily.

so in the long run, WAR really only gets somewhere between 10-15% persistent mitigation. PLD always gets its 20%. Not only that, PLDs secondary forms of mitigation are at 0 cost. they don't take MP, they dont' take TP, they don't even consume the GCD.

Now you might argue that WAR has self healing to help bridge this gap. but keep in mind, we're still trying to bridge the gap between persistant mitigation. we haven't even touched the extra bonuses paladin gets between shield, parry, rampart, bulwark, and sentinel. WAR will need to spend the bulk of its active abilities simply trying bridge that persistent 20%. The mitigation gap is just too large. no amount of 'research' is going to bridge that gap.

-------------------------------------------

Now, there are a couple of abilities I rarely use on War. and so in the short term this will be where I focus my efforts.

1) Unchained

I never use this. I can't really see a point to it. you can't use it as an 'o shit' button like IB. you could potentially use it with berserk/blood bath/vengeance to get more health returned. but then you're losing the 15% healing received. I imagine this breaks even, or is close enough to not matter, in the long run. but I could be wrong. So i'll test it out in my own time. I have a feeling this is a lot like Sword Oath. it's an ability we aren't really meant to use that often.

2) Holmgang.

I can think of 0 situations where I would even attempt to use this. given how mobile you have to be to dodge attacks, why in the world would I ever bind myself? is this a potential hidden gem? if I bind myself can I avoid knockbacks like Repel? is that even really useful? when a mob does one if it's 'don't attack the tank' moves can I bind it and interrupt these abilities? I mean it doesn't even do damage... I suspect this is WARs answer to Cover. but I can't even bind something to run up and get hate back. Coupled with the high probability of this getting you killed...I'm pretty sure this isn't what we're missing either.

3) Stat allocation.

I wont' dive deep into this. it's fairly obvious. However all data I've seen has been inconclusive. that being said, I doubt us switching from VIT to STR builds or going with DD gear and attempting to tank with it will drastically improve our ability to stay alive. not to mention it's not intuitive to do this. if this is the intended stat allocation, its not clear enough. therefore it is bad design.

[Ehayte]

1) if it's taken the entire playerbase this long to figure out, it's too well hidden. Therefore it is a design problem.

This is kind of the smartest and most on-topic thing anyone has said in this thread. Either Warriors scale way better with gear (like, get way bigger advantages with gearing up because PLD is on a % mitigation where Warrior uses hard numbers) and they are testing with better gear, or they need to share with us how to play this class.

After participating in these threads, etc, I really would like to switch classes to a WAR because of how interesting it sounds, but it just seems like the odds are insurmountable currently. WHETHER its as an off-tank or a main tank.

[Sargon]

so in the long run, WAR really only gets somewhere between 10-15% persistent mitigation. PLD always gets its 20%

Just a reminder, heal buff % cannot be simply compared to damage reduction.
100 -20% = 80

10 hits => 800dmg on PLD, 1000dmg on war
+15% on a 800hp heal = 920hp

To have those 1k hp healed back you need a permanent 25% heal buff.

[Ruminate]

Either Warriors scale way better with gear (like, get way bigger advantages with gearing up because PLD is on a % mitigation where Warrior uses hard numbers)

Why do so many people keep saying this?
All percentage multipliers inherently scale. Doesn't matter if the percentage multiplier is vitality or damage reduction, both scale.

[Kyzr]

Just a reminder, heal buff % cannot be simply compared to damage reduction.
100 -20% = 80

10 hits => 800dmg on PLD, 1000dmg on war
+15% on a 800hp heal = 920hp

To have those 1k hp healed back you need a permanent 25% heal buff.

^This.

Who cares if I have 25% more HP if it still takes 20% more mana and GCD to heal through the damage we're taking? Even if you have a BRD and the mana never becomes an issue, a healer still has to heal a WAR 20% more OFTEN than a PLD. There's no permanent +20% casting speed buff in the game, and even if there was, PLD would still be leaps and bounds ahead. Why? Because there's no way a WARs damage is better than 2x Healers doing damage instead of 20% more healing.

For a WAR to be better than PLD, they have to do more damage than the PLD plus the extra damage healers get to do when they have a PLD. There's no way that's the current state, nor should it ever be.

[Delphineas]

Didn't FFXI have a boss named Absolute Virtue? Didn't SE say they beat him internally? Wasn't he undefeated for 2 years without "exploits"?

And people really believe the dev when he says wars are fine?

*never played eleven, just heard about it.

[Kitru]

Kitru.. How much damage does Sword Oath can do? Its 50 but not 50%.. And that's only applicable to auto attack (Correct me if im wrong).

Sword Oath does an extra 50 damage with each auto-attack, which happens more than once per GCD so the baseline estimate I was going with (one hit per GCD) is actually underestimating PLD DPS in Sword Oath.

You are right on berserk though but how does maim skill this advantage "disappear" out from a warrior?

Because you're acting as if PLD and WAR have the same base damage before multipliers, which they do not. A WAR has a fair deal *less* damage pre-multipliers because PLD has Circle of Scorn and Spirits Within.

War does and able to do more damage than a PLD. It's depending on how you are going to play with it, combining what kind of based class skills & cross class skills before entering a dungeon.. Too bad you cant proved that..

I've already proven it elsewhere. Let me prove it, once again, right here and now. To clarify, we're talking about WAR outside of Defiance and PLD in Sword Oath. These are *not* tank active numbers. Furthermore, there is no difference between the damage that a sword and an axe deal. The only difference *at all* is that Sword Oath does more damage with a faster weapon and, as previously stated, I'm acting as if the weapon speed for it were 2.5 for simplicity's sake, which ends up underestimating PLD DPS by a small amount.

The only additional that is going to improve a WAR's DPS is Internal Release. It provides 20% additional crit for 15 seconds every 60 seconds. With a baseline 10% crit rate and 50% additional damage from crits, this amounts to a 9.5% ((1+3*.5)/(1+.1*.5)) increase in damage with 25% uptime (15/60), which equates to a monumental 2.375% (9.5 * .25) increase in damage dealt.

Berserk has separate benefits for auto-attacks and special attacks thanks to pacification. For auto attacks, Berserk provides 50% additional damage for 20 seconds every 90 seconds, or 11.1% increase to damage (1/.9, yet again). For special attacks, it provides a 50% increase for 20 seconds and a -100% debuff for 5 seconds every 90 seconds. Effectively, it's got a 10 second duration so that amounts to 5.56% increase to damage (.5/.9).

Maim is a 20% increase in damage on a permanent basis. Storm's Eye is a ~11.1% (1/.9) increase in damage dealt, but it applies to both PLD and WAR if there is a WAR present.

Fight or Flight provides a 30% increase in damage dealt for 30 seconds every 90 seconds, which amounts to a 10% increase in damage dealt over time. Spirits Within deals 300 potency every 30 seconds (as DPS, you shouldn't be taking damage), and Circle of Scorn provides 250 potency every 25 seconds. As such, that's an extra 25 damage per GCD (300/30*2.5; 250/25*2.5) since they're both off GCD.

Fracture provides a DPS increase by providing a 220 potency attack every 3 Halone uses. This necessitates some downtime on the debuff, but it's required to prevent interrupting a combo or clipping ticks. For a WAR, it's 300, but the optimal rotation clips one tick so it's practically only 280.

Auto-attacks provides a baseline 83.33 potency per GCD (2.5/3). Sword Oath provides 50 additional potency with every auto attack hit and swords tend to have attack speeds in the 2.1-2.3 range, but, as previously stated, I'm simplifying which underestimates performance slightly.

The effective rotations for each are as follows:

PLD: Halone>Halone>Halone>Frac (150 + 200 + 260 + 150 + 200 + 260 + 150 + 200 + 260 + 220) = 205 per GCD
WAR: SE>BB>Frac>SE>BB (150 + 190 + 270 + 150 + 200 + 280 + 300 + 150 + 190 + 270 + 150 + 200 + 280) = 213.85 per GCD

The total damage formulas are then...

WAR:
(213.85 * 1.0556 + 83.33 * 1.11) * 1.2 * /.9 * 1.02375 = 434.39 potency per GCD

PLD (without WAR):
(205 + 83.33 + 50 + 25 + 25) * 1.1 = 427.16 potency per GCD

PLD (with WAR):
(205 + 83.33 + 50 + 25 + 25) * 1.1 / .9 = 474.63 potency per GCD

So, yeah, PLD in Sword Oath deals only slightly *less* damage than WAR outside of Defiance *and that's underestimating PLD damage*. WAR has a lot of damage modifiers but it has a lower base value to act upon. This is the issue that people keep forgetting. It's not like PLD and WAR start off with the same base damage and WAR ends up doing more thanks to a crapton of amazing multipliers. PLD ends up doing more because Fight or Flight is flipping awesome, Berserk is terrible, and Circle of Scorn and Spirits Within are, effectively, a 14.78% increase in total DPS (it's even larger in Shield Oath because it doesn't have that pesky 50+ additional getting in the way).

The only reason a PLD will deal less damage than a WAR is if said PLD isn't using the tools given to them like they should be. If they're not using Circle of Scorn and Spirits Within, their DPS plummets (372 potency per GCD).

Now, because I know you're going to ask for it, here are the numbers while in their relevant tank stances. Everything is the same except that they have their relevant damage multipliers (.75 for Defiance, .8 for Shield Oath), PLD no longer gets the +50 potency, and WAR gets a 4.76% increase to DPS from Wrath V ((1+.2*.5)/(1+.1*.5)). Internal Release actually drops its contribution to 2.72% increase (((1 + .4 * .5)/(1+.2 * .5) - 1)*(15 / 60)).

(Edit: I'm adding a pure Butcher's Block, max enmity generation formula; base DPS per GCD is 210: (150 + 200 + 280) / 3)

(Edit: Butcher's Block is actually inferior to a new combo rotation that I've been playing with: BB>BB>SE; the math on it is slightly more complex because you're sacrificing 100% uptime on SE, but I'm adding it as well; the damage numbers are 630 for a BB combo and 610 for SE)

WAR:
(213.85 * 1.0556 + 83.33 * 1.11) * 1.2 /.9 * 1.0476 * 1.0272 * .75 = 342.45 potency per GCD

WAR (Butcher's Block):
(210 * 1.0556 + 83.33 * 1.11) * 1.0476 * 1.0272 * .75 = 253.56 potency per GCD

WAR (BB>BB>SE):
((2 * (630) / .9 + 610) / 9) * 1.0556 + 83.33 * 1.11) * 1.2 * 1.0476 * 1.0272 * .75 = 325.3 potency per GCD

PLD (without WAR):
(205 + 83.33 + 25 + 25) * 1.1 * .8 = 297.73 potency per GCD

PLD (with WAR)
(205 + 83.33 + 25 + 25) * 1.1 * .8 / .9 = 330.8 potency per GCD

A PLD without a WAR around to provide the debuff deals 86.9% of the damage that a WAR does, but as soon as you go WAR + PLD, the PLD gets pretty damned close (96.60%). Of course, as soon as you start not using Inner Release and Berserk on CD in order to maximize mitigation through Inner Beast, the WAR number starts going down pretty quickly. A PLD can use his CDs as quickly as possible without negatively impacting anything.

(Edit: By maximizing enmity generation using the BB>BB>SE combo, WAR damage drops a fair deal, reducing the DPS advantage. With 2 WARs, you can do some really dangerous things if you know how to cycle SE combos properly to maintain 100% uptime while using BB>BB>BB>SE>Fracture, but that requires the WAR synchronize adding in a Fracture use would increase damage but creates some problems with maintaining uptime and combos. If you do it right, you actually maintain *better* damage than a WAR solo.)

Basically, there is no DPS advantage to using 2 WARs (Edit: You *can* get a DPS advantage out of running 2 WARs, but it is *really* hard to manage and requires a lot of precision between both players). To max out damage, comically enough, you want a PLD offtank and a WAR main tank since PLD does a crapton more damage with Sword Oath than a WAR does without Defiance. Of course, the maximization of said damage is actually a comparatively small amount of damage. The difference between a PLD tank and a WAR tank exclusive of one another, is a whopping 31 potency per GCD. When you consider that each DPS is going to do in excess of twice the damage that a tank will deal (try holding aggro with just the Storm's Eye combo and Defiance; Defiance has a 2x enmity modifier so you should be able to match/compete for aggro unless DPS does a crapton more) and that there are twice as many DPS as tanks, that 40 DPS amounts to a whole 2.67% (40/(300+600+600) increase in total damage dealt. In an 8 man group, the difference is even smaller because you've got twice as much DPS.

Furthermore, here's the math for enmity generation.

Tank stances double enmity generation. Crits do not generate any extra enmity beyond that caused by their increased damage.
Halone generates 2050 enmity per cycle pre-tank stance (150 + 200 * 3 + 260 * 5). Put that into the optimized combo and you get 637 enmity per GCD ((2050 * 3 + 220) / 10) before multipliers and additional potency.
Butcher's Block generates 2150 enmity per cycle pre-tank stance (150 + 200 * 3 + 280 * 5). Storm's Eye generates only 610 (150 + 190 + 270). Put that into the optimized combo and you get 446.15 enmity per GCD ((2150 * 2 +610 * 2 + 280)/13)). (Edit: I'll also include a enmity generation numbers for pure Butcher's Block spam, which is 716.67 enmity per GCD)

(Edit: adding BB>BB>SE numbers)

WAR:
(446.15 * 1.0556 + 83.33 * 1.11) * 1.2 /.9 * (1 + .0476) * (1 + .0272) * .75 * 2 = 1212.66 enmity potency per GCD

WAR (Butcher's Block only):
(716.67 * 1.0556 + 83.33 * 1.11) * (1 + .0476) * (1 + .0272) * .75 * 2 = 1370.43 enmity potency per GCD

WAR (BB>BB>SE):
(((2 * (150 + 600 + 1400) / .9 + 610) / 9) * 1.0556 + 83.33 * 1.11) * 1.2 * 1.0476 * 1.0272 * .75 * 2 = 1403.18 enmity potency per GCD

PLD (without WAR):
(637 + 83.33 + 25 + 25) * 1.1 * .8 * 2 = 1355.78 enmity potency per GCD

PLD (with WAR):
(637 + 83.33 + 25 + 25) * 1.1 * .8 * 2 / .9 = 1506.70 enmity potency per GCD

Pay no attention to the italicized. It is conclusions derived from from old numbers that I'm leaving for historical perspective.
WARs *do* generate a metric shitton more enmity than a PLD does (this actually surprised me; thanks to the difference in spamming Halone v. alternating BB and SE, I fully expected PLD to win out handily, but the PLD "bonuses" are flat value whereas the WAR bonuses are multipliers so, because baseline enmity generation is higher for both, the bonuses that PLD gets are *heavily* diluted). This is the *only* advantage that WAR has. Of course, a PLD is more than capable of maintaining all of the threat that it might need and redundant enmity is completely pointless: there's no difference between having 1 billion more enmity than the DPS and having 1 more enmity; either way, you have aggro. As such, the "advantage" is less important that it might otherwise be.

Editing in new conclusions:
Apparently my initial belief that PLD is better at enmity generation is supported: PLD *does* generate better enmity per GCD because it gets to spam its high enmity combo ad nauseum. This just makes WAR look even worse: you have to choose between slightly better damage but dramatically worse enmity generation, likely enough to have really good DPS riding your ass nice and hard, or enmity generation that just *barely* edges out PLD along with utterly abysmal DPS. It would be interesting to get a conversion rate between DPS potency value and tank potency value to determine what the relative enmity generation between the two is. As it stands, we can only guess at whether WAR actually has enough to keep aggro off of a mathematically perfect DPS.

Edit after adding BB>BB>SE:
WAR can manage to edge out PLD in solo enmity generation. A PLD with a WAR is still going to match DPS while generating slightly better damage. Really, nothing really changes about the end conclusions since the two classes are still effectively identical. Pointing out the fact that the differences between the two classes are minute, which was the point of all of this, is still completely valid.

P.S. Previous maths I did didn't account for Wrath V and Internal Release, hence the discrepancy with those numbers.
P.S.S Yes, I recognize that I'm reversing my position on enmity generation between the two, but I've got no problem changing my view based upon math and evidence.
P.S.S.S Holy crap, all of that math was fun.
P.S.S.S.S After further revision and correction of numbers, apparently my initial view on enmity generation were correct. No, I'm not flip flopping. I follow the numbers, and we're continually honing them.

[ZephyrK]

I'm starting to think we need a Yoshi P's facts website!

Please wait for it.